Subevents detection through topic modeling in social media posts

Nolasco, Diogo; Oliveira, Jonice

doi:10.1016/j.future.2018.09.008

Cited by 29 publications

(18 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Text mining provides semi-automated methods for mining available facts. Some of the applications of text mining include topic modelling (Nolasco and Oliveira, 2018; Li et al , 2018), sentiment analysis (Saleena, 2018; Xiaomei et al , 2018), text categorization (Sboev et al , 2016; Park and Kremer, 2017) and document summarization (Chaix et al , 2018; Eskici and Koçak, 2018).…”

Section: Methodsmentioning

confidence: 99%

Using the fuzzy multicriteria decision making approach to evaluate brand equity: a study of privatized firms

Dınçer

Bozaykut-Buk

Emir

et al. 2019

JPBM

View full text Add to dashboard Cite

Purpose The purpose of this paper is to present a multidimensional evaluation of brand equity performance incorporating dimensions adopted from the balance scorecard (BSC) approach to business performance. Design/methodology/approach In this study, text mining is used for automatic extraction of valuable information from textual data such as the financial reports of firms. Instead of expert opinions, linguistic scales built upon outcomes of text mining are used as inputs for decision-making. The proposed model combines fuzzy DEMATEL (FDEMATEL), fuzzy ANP (FANP), fuzzy TOPSIS (FTOPSIS) and fuzzy VIKOR (FVIKOR) methods for weighting criteria and ranking alternatives. Findings Using data from five privatized firms in Turkey, the study’s findings demonstrate that the customer is the most important dimension of brand equity performance evaluation. Cash flow and brand loyalty are identified as the most important criteria in the measurement of brand equity performance. Practical implications Findings highlight the importance of firms taking action to increase consumer perceptions, attitudes and behaviors in the privatization processes. For this purpose, privatized firms need to understand the expectations of customers to increase customer satisfaction and loyalty and therefore improve brand equity. Originality/value The paper contributes to literature in several important ways. First, by adopting the BSC approach, it proposes a holistic and a multidimensional model for measuring brand equity performance. Second, the study offers a novel methodology using a hybrid multi-criteria decision-making model designed for the fuzzy environment. Third, the study uses the knowledge extraction tool of text mining in the fuzzy decision-making process. Finally, the study evaluates the brand equity performance of privatized firms in an emerging country context.

show abstract

Section: Methodsmentioning

confidence: 99%

Using the fuzzy multicriteria decision making approach to evaluate brand equity: a study of privatized firms

Dınçer

Bozaykut-Buk

Emir

et al. 2019

JPBM

View full text Add to dashboard Cite

show abstract

“…The techniques used for Topic Detection and Labeling are an implementation of the ones described by [Nolasco and Oliveira 2018], as they prove successful and efficient when applied in the dataset types used in our experiments, namely, in the academic and social network domains. The Topic Correlation approach is based on the Kullback-Leibler divergence [Kullback and Leibler 1951], a measure used to associate two probabilistic distributions and that we use on the distributions resulted from topics models.…”

Section: Proposalmentioning

confidence: 99%

“…In this task, we are using the Latent Dirichlet Allocation (LDA) to detect topics from the textual collections. For this, we also use the topic detection methods described in [Nolasco and Oliveira 2018].…”

Section: Topic Detectionmentioning

confidence: 99%

Topical Rumor Detection based on Social Network Topic Models Relationship

Nolasco

Oliveira

2021

iSys

Self Cite

View full text Add to dashboard Cite

The rumor detection problem on social networks has attracted considerable attention in recent years with the rise of concerns about fake news and disinformation. Most previous works focused on detecting rumors by individual messages, classifying whether a post or blog entry is considered a rumor or not. This paper proposes a method for rumor detection on topic-level that identifies whether a social topic related to a reference or authoritative topic is a rumor. We propose the use of a topic model method on social, scientific and political domains and correlate the topics found to detect the most prone to be rumors. Two scenarios were analyzed; the Zika epidemic scenario where our reference set of topics are scientific and the Brazilian presidential speeches where our reference set is extracted from the political speeches themselves. Results applied in the Zika epidemic scenario show evidence that the least correlated topics contain a mix of rumors and local community discussions. The Brazilian presidential speeches scenario suggests a strong correlation between rumor topics from both the speeches and the social domains.

show abstract

“…Currently, Latent Dirichlet Allocation (LDA) introduced by Blei et al (2003) represents one of the most widespread approaches. Despite the fact that these methods have traditionally been closely associated with the analysis of textual datafor example, consumer/customer feedback (Bastani et al, 2019;Hu et al, 2019), social media content (Curiskis et al, 2020;Nolasco and Oliveira, 2019) or research interests (Xiong et al, 2019;Yang et al, 2019) -LDA has also already been successfully used in different biological settings, particularly in genomics (Chen et al, 2010;Perina et al, 2010;Pratanwanich and Lio, 2014;Shiraishi et al, 2015;Yu et al, 2014;Zhang et al, 2012). However, to the best of the authors' knowledge, its prospects within food science still remain to be elucidated.…”

Section: Introductionmentioning

confidence: 99%